Real-time marketing measurement with multi-touch attribution for CPG

Reading Time: 6 minutes

5 Best Practices for Deploying Machine Learning Models

Digital marketers are looking for near-real-time intelligence on each customer touchpoint to optimize campaigns and create personalized customer experiences. Multi-touch attribution (MTA) empowers marketers with the data across customer journeys to identify the value each touchpoint brings to drive conversion and in turn boost marketing tactics at a granular level.


A recent podcast conducted by The CPG Guys led by Sri Rajagopalan (Chief Customer Officer at General Mills) in conversation with Taj Peeran (VP Marketing of Digital, eCommerce, and Brand Engagement at Reckitt) and Rahul Kumar Singh (Co-Founder and Chief Analytics Officer at Sigmoid) discussed MTA, hyper-personalization, CLTV, and the role of AI in creating business value for the CPG industry.

Podcast excerpts

Sri Rajagopalan (Moderator): What was your inspiration to start Sigmoid? Tell us about the data science practice and how you have been delivering ‘measurable outcomes’ for CPG customers?


Rahul Singh: I founded Sigmoid in 2013, with two of my IIT batchmates Lokesh and Mayur, to help companies make faster and smarter business decisions using data and AI. In the last nine years, we have helped several companies build custom data solutions to solve complex business challenges.


Podcast by Rahul Singh


We are helping them achieve success in business initiatives such as personalization, demand forecasting, marketing measurement, and more.


Sri Rajagopalan: How does Sigmoid’s Multi-Touch Attribution accelerator help CPG companies optimize marketing spends and drive revenue?


Rahul Singh: CPG companies are looking for reliable ways to map their digital investments, especially as we’re heading towards a cookieless world and the advent of customer privacy laws like GDPR and CCPA. Another aspect is spend optimization, where CPG companies spend a large amount of investments to drive a business goal like measuring revenue, viewability, reach impressions and so on. This is where Sigmoid’s customizable Multi-Touch Attribution accelerator helps the marketers in the following ways:


  • It is a productized service that allows marketers and brand managers to optimize their digital advertising spends by generating the required insights.
  • It enables inflight campaign optimization, which is a core requirement of marketers.
  • It provides a faster, more accurate dashboard that can measure marketing tactics at a granular level allowing marketers to measure their campaigns within one to two weeks of performance.
  • Its quasi-experimentation design can deliver on speed, cost and scale at the same time.


Sri Rajagopalan: Can you connect Sigmoid’s capabilities directly with Reckitt’s eCommerce results?


Taj Peeran:
Sigmoid helps us normalize all the campaign data into a single place. The team provides us with access to real-time dashboards that basically break down channel performance at a creative level, mark the audience, plan our investments and identify the aspects that are really working for us. On the e-commerce front, we examine Amazon sales data that are not attributed to the media, to understand how it is impacting our organic sales. We can then optimize overall media investment to generate both offline and online sales by building relevancy and achieving organic sales. Sigmoid’s model helps us generate bi-weekly insights in a cost efficient manner to visualize the impact of activation. In this way, we’re able to successfully measure the impact of our digital campaigns on offline sales.


Sri Rajagopalan: What does AI mean in the CPG industry from an ‘outcome’ perspective and how does Sigmoid leverage AI?


Rahul Singh: Over the years, AI is adding value to the CPG industry and is bringing new initiatives that power insights into promotions, marketing, stocks, distribution, and more. Having accurate and granular AI models can drive incremental revenue growth by more than 10%.


Another core area where AI drives significant value is creating hyper personalized consumer experiences such as pricing or promotions. We have built an AI powered consumer intelligence engine that constantly ingests real time data and measures metadata related to consumer behavior, decision making, preferences, and interests to generate relevant recommendations to maximize value for businesses.


In addition, we have developed capabilities to build intelligent models at a single consumer level that get self-optimized to learn about consumers and identify most appropriate marketing techniques.


Taj Peeran: The basic principle of AI and machine learning is to become significant business drivers. AI models like that of Sigmoid can help deliver hyper-personalized consumer experiences for the website recommending similar products to upsell or cross sell. Additionally, CPG companies are starting to realize the benefits of creative optimization where machine learning algorithms help deliver dynamic creatives or ads based on consumer’s interaction with other elements, in real time without manual intervention, which is quite innovative.


Sri Rajagopalan: How are you driving value for your clients in Customer Life-Time Value (CLTV) and Hyper-Personalization?


Rahul Singh: Personalized recommendations that drive better Customer Life-Time Value (CLTV) is an important initiative for every consumer-focused brand to retain and improve their top line growth. CLTV uses a lot of different approaches and personalized recommendations. Personalized recommendations can be of two types— customizing recommendations at a segment-level and treating a single consumer as a unique individual. We have carried out personalized recommendations with multiple Fortune 500 companies catering to both approaches, depending on their level of maturity and availability of either the data or the channels that they are targeting. In one such instance, we built a robust system for a global Fortune 500 company to improve their CLTV by defining the ML models and then understanding the current and future potential of a consumer’s value. We helped them identify consumers as high potential to build stronger and better relationships to improve their lifetime value.


Rahul Singh Podcast


Sri Rajagopalan: What did you find different and compelling in Sigmoid’s approach? And why did you pick them vs other analytics companies?


Taj Peeran: Sigmoid has been a preferred partner for a number of initiatives including data analytics and machine learning. And there are three clear ways they’ve added value for us.


  1. Problem Solving: Sigmoid brings a unique blend of data science, big data engineering, and business consulting to build self learning decision pipelines on cloud.
    They have an innovative approach to problem-solving owing to the fact that their teams are highly invested in our business problems to deliver tangible business results.

  3. Transparency: What we also like about this partnership is transparency that they ensure into machine learning models. Unlike other vendors that provide black-box solutions, Sigmoid ensures that our team is fully aware of the algorithms behind the white-box data solutions, which is a significant differentiator.

  5. Increased ROI: The speed to market and the cost efficiency they are able to bring to maximize ROI is very enterprising.

I would strongly recommend Sigmoid to organizations for its flexibility and commitment to solve real business challenges.


Sri Rajagopalan: Rahul, what is your value proposition that differentiates you? And why should a brand work with you versus somebody else?


Rahul: We have a proven experience in delivering high business value for our clients with a team driven toward achieving business goals. There are four key factors that we focus on to achieve customer’s goals.


  1. Defining success: Primarily, we define the success for a particular project in an objective manner to assign a benchmark that we need to achieve. From a machine learning perspective, we define analytical measurement KPIs that directly replicate the business goal.

  3. Faster Innovation cycle: We build infrastructure to achieve fast innovation cycles to bring high quality data within one hour or a few minutes from 10-15 days.

  5. High impact business solutions: We focus on bringing high value for the business in a consistent and reliable manner.

  7. Talent: We have hired the top talent from India and abroad with a very efficient training process. In addition, we are the largest data platform in the world processing more than 300 terabyte data. This is backed by our strong engineering capabilities, which helps us on the data science side.

About the author

Khyati Dubal is a Senior Associate, Content Marketing at Sigmoid. She has over six years of experience in writing on advanced technologies that power the growth of next-generation tech companies.

Transform data into real-world outcomes with us.